# IDK Complex Hairstyles V2

## About

Specialized in intricate and detailed hairstyles — braids, buns, ringlets, twintails, and more. If the base model gives you "generic hair blob", this fixes that.

PT-BR: LoRA focada no conceito 'Complex Hairstyles' para modelos SDXL baseados em Illustrious. Reforca a geracao de dreadlocks, drill hair, twintails.

## Trigger words

### Primary triggers

```

dreadlocks, drill hair, twintails

```

### Supporting tags

Frequent thematic tags from the training set - combine them with the primary triggers to steer the concept:

```

blonde hair, long hair, hair ornament, short hair, black hair, red hair, pink hair, hair ribbon, blue hair, short twintails, hair bow, twin drills, magical girl, yellow eyes

```

Also works with the V1 triggers: french braids, dreadlocks, cornrows, space buns, messy bun, drill hair, ojou ringlets, twintails, ponytail

## Example prompts

```

masterpiece, best quality, absurdres, woman with intricate braided updo, long flowing, wearing delicate accessories, princess, detailed strands, <lora:IDK_complex-hairstyles_V2:0.9>

```

```

masterpiece, best quality, absurdres, woman with elaborate fantasy hair sculpture, elaborate curls, hair flowing wildly, seductive, detailed strands, <lora:IDK_complex-hairstyles_V2:0.9>

```

```

masterpiece, best quality, absurdres, woman with messy elegant waves, long flowing, wearing delicate accessories, princess, detailed strands, <lora:IDK_complex-hairstyles_V2:0.9>

```

## Preview generation settings

| Setting | Value |

| --- | --- |

| Checkpoint | IDN_Illustrious_V10_T |

| Size | 1280x720 + hires 1.5x RealESRGAN 4x+) = 1920x1080 |

| Sampler | Euler a |

| Steps | 28 (+14 hires) |

| CFG scale | 5.5 |

| Denoising strength | 0.4 |

| Seeds | 1000-1009 |

Negative prompt: worst quality, low quality, bad anatomy, bad hands, extra digits, fewer digits, missing fingers, jpeg artifacts, signature, watermark, username, blurry, lowres, bad proportions, deformed, mutated

## Recommended weight

Use 0.8-1.1. This file is a V1+V2 merge (rank concatenation), which slightly dilutes each source - it tolerates and often benefits from weights a bit above 1.0.

## Changelog V2

- New, larger and more diverse training dataset.

- 1080p preview images.

- Merged with the V1 LoRA, so V1 behavior is preserved.

Base model: Illustrious SDXL IDN_Illustrious_V10_T)

Training pipeline: [IDK-Trainer](https://github.com/NOTcisLol/IDK-Trainer)

File: IDK_complex-hairstyles_V2.safetensors